This paper provides an empirical analysis of caste discrimination in the regular urban labour market in India. The affirmative action policy is confined to the minuscule public sector and excludes the vast private sector; therefore, analysis of caste discrimination has been conducted separately for public and private sector workers. To examine the wage gap between workers of forward castes (others) and lower castes (Scheduled Castes), the 50th, 61st, and 68th rounds of the Employment and Unemployment Survey data of the National Sample Survey Organisation (NSSO) have been used. The main conclusions, based on decomposition analysis, are that (a) endowment difference contributes more than discrimination to the raw wage gap; so, expanding educational opportunities for Scheduled Castes can be a useful strategy to reduce discriminatory treatment against them; (b) wages are lower for Scheduled Castes than for equally qualified forward castes by 19.4 per cent in the public sector and by 31.7 per cent in the private sector; (c) occupational discrimination, or unequal access to jobs, is more important than wage discrimination in both public and private sectors; and (d) the quantile regression results reveal a “glass ceiling effect” in the private sector and a “sticky floor effect” in the public sector. The empirical findings provide strong evidence for the extension of affirmative action policy in the private sector.